Machine Vision

cancel
Showing results for 
Search instead for 
Did you mean: 

Invalid Pixel Format for JAI AT200GE

Solved!
Go to solution

Hi,

 

I'm trying to use an JAI AT200GE camera which contains 3CCDs and a prism to split the colours out to each one. The camera is capable of outputting 8bit RGB and two 10bit packed formats RGB10V1Packed and RGB10V2Packed (which is now called RGB10P32 according to the latest picel format naming convention). The extra bit depth is pretty important in my application but I cannot find a way to get LabVIEW/MAX to decode the 10 bit pixel formats.

 

As none of these are standard in MAX, is there anything I can do?

 

Really appreciate any help.

0 Kudos
Message 1 of 6
(3,530 Views)
Solution
Accepted by topic author markquinn

What is the exact name used by the RGB10 packed formats?

 

The latest version of IMAQdx has support for the legacy formats "RGB10PackedV1", "RGB10PackedV2", as well as the new PFNC variant "RGB10p32" (which is identical to RGB10PackedV2). As long as the camera names the formats as one of these, it should work in IMAQdx.

 

If they use a slight variation of the name, that could be an issue. You could edit the camera's XML file in the local IMAQdx cache directory and change the names to the appropriate ones above as a workaround.


Eric

0 Kudos
Message 2 of 6
(3,509 Views)
Solution
Accepted by topic author markquinn

@BlueCheese stole my thunder, but for future reference, we have a list of supported pixel formats here.

 

And if you don't know where the local IMAQdx cache directory is, it should be in a path similar to this: 

C:\Users\Public\Documents\National Instruments\NI-IMAQdx\Data

 

In the XML/ subdirectory, you can hand edit the XML file and IMAQdx will use that modified one instead of re-downloading it from the camera.

Message 3 of 6
(3,506 Views)

Thanks for the input and speedy reply! I'll have a go with that tomorrow (Friday = Lab Day) and let you know how I get on. 

 

It's a bit of a strange camera this one because it has three CCDs in it and a prism to split the colours out there is no need to do any Bayer decoding or anything like that... I will update soon.

0 Kudos
Message 4 of 6
(3,476 Views)
Solution
Accepted by topic author markquinn

I have just filed a bug report about the pixel format naming. According to the GenICam PFNC, it should be RGB10V1Packed, as the camera uses.

BR / Ole K.J.

0 Kudos
Message 5 of 6
(3,456 Views)

I had a really helpful teamviewer session with the support from JAI this morning and it turns out that this camera behaves much better with the windows firewall off.

 

In order to get the pixel decoding correct you need to change the XML file that comes from the camera. If you change all instances of RGB10V1Packed to RGB10PackedV1 everything works perfectly.

 

Thanks so much for the advice and many thanks to Ole from JAI.

0 Kudos
Message 6 of 6
(3,449 Views)